Stake to fund the next material discovery. Research plan in. Lab protocol out.
Eight specialized material science ANNs are live in the Aigarth marketplace today. A research question becomes a 5-stage plan, a DFT-validated candidate, and a printable lab protocol. Stake QUBIC to participate in the network, and in the revenue when a discovery ships.
The team
Eight small, specialized ANNs collaborate to discover and validate a new material. Each is published, versioned, rateable, and live in the Aigarth marketplace today.
Plan a material discovery workflow from a research question.
Ingest scientific papers into a structured knowledge graph.
Run DFT, MD, or ML surrogate simulations for material properties.
Sanity-check a simulation result against first principles.
Generate candidate materials for a target property profile.
Pareto-sort candidates on (energy density, cost, cycle life).
Convert a candidate material into a synthesizable lab protocol.
Cross-check a prediction against literature and historical data.
All 8 are running on the Aigarth stub backend. Every deploy, stake, and call is recorded. The format-only signature on each version is the Phase 0 placeholder: real K12 signatures ship once the Qu bic verification milestone lands.
The math
Real numbers from the Phase 17 architecture evaluation. Every cost and every revenue share is visible before you commit.
DFT refinement dominates at 87% of the total.
A monthly reservation, billed per workflow.
Stakers get 30%. A 1,000 QU/month stake earns 0.3 × network revenue.
Numbers are illustrative Phase 0 placeholders. They will be updated as we ship and measure. The cost structure matters more than the number. A researcher can always see the per-stage breakdown before committing.
How a discovery flows
A research question becomes a plan. Each stage consumes the plan and produces an artifact. The Physics Reasoner in the middle decides whether to flag a stage for re-run.
Every stage emits a structured prediction with uncertainty. A bad stage can be re-run alone. The rest of the workflow is preserved.
The Director's plan includes a per-stage cost + duration. The user must confirm before the workflow starts. No $1,000 surprise at the end.
The Simulation Runner routes through a fast ML surrogate (MACE) before paying for full DFT. Cuts wall time by 10× and cost by ~9×.
The same simulation is never paid for twice. The orchestrator hashes the input deck and serves from cache.
Real attribution, on chain
When a material discovery is published, the on-chain attribution is fixed immediately. That's what makes it useful as a real citation in a real research paper.
A new Qubic primitive that ships in Phase 4. The publish action is one-way and immediate: not via daily rollup. Once a material is published, the ANN authors, the workers who ran the simulations, the researchers who funded the work, and the stakers who backed the network are all recorded. Immutable. Citable.
The same primitive will be reusable for other high-stakes use cases : drug discovery, clinical trials, regulatory submissions. The architecture stands to benefit more than just material science.
Why this fits the existing Aigarth stack
258 story points and 7 backend services already ship. The 8 material science ANNs run on the same infrastructure that powers the rest of the marketplace. No new architecture.
services/ann: registry, versioning, reviews, deploy
All 8 ANNs are published, versioned, rateable, and deployable as compute jobs. Same as the other 6 ANNs in the marketplace today.
services/marketplace: listings, offers, auctions
License the whole 8-ANN workflow as a single pre-composed 'Battery Cathode Discovery Pack' listing. Pay per call, or per workflow.
services/gateway: LLM call routing
Director, Literature, Experiment Planner route through the OpenAI-compatible gateway. Real LLM today, stub during Phase 0.
services/qubic: Qearn staking, on-chain registry
Stake QUBIC on Qubic → Aigarth credit balance → per-workflow deduction. Revenue share flows back to stakers via Qearn.
Stake to participate
Four ways to join. Most material science researchers and computational labs start with the Builder or Infrastructure Partner tier.
Foundational staker. The earliest way to participate.
For people who want to be part of the founding moment.
Developer / researcher. Earn allocation, deploy material science ANNs, get marketplace priority.
For people who will build and run on the network.
Node operator. Register a DFT or MD worker, earn per simulation.
For HPC labs with idle compute.
Organization. Private infrastructure, dedicated support, custom SLAs.
For research institutions and corporate R&D.
All four tiers run through the Aigarth Genesis Offering at qubic.org/genesis. Material science participants get marketplace priority placement and 0% commission on the first 1,000 workflows (Phase 0 promo).
What ships next
5 phases, 84 story points estimated total. Phase 0 is done: 8 ANNs live as stubs. Re-evaluation gate at the end of Phase 1.
8 specialized ANNs registered, listed in the marketplace, deployable as jobs. Real LLM, synthetic compute.
Ingest 1,000 papers, answer a research question in < 5 min. Director + Literature real. Rest still stubs.
DFT, MD, MLIP workers. End-to-end material discovery. ~52 QU per workflow, ~102 h on 1× CPU.
Per-ANN quality ranking, 'Battery Cathode Discovery Pack' pipeline listing, reviews + ratings.
On-chain ANN ownership, worker registry, reputation, rewards, and the new discovery_attribution contract.
Stake. Train. Discover.
The 8 ANNs are live today on stubs. The Phase 1 knowledge prototype is the first real gate. Stake to be part of the founding moment, and to participate in the revenue when the first discovery ships.
This page is part of the Aigarth Cloud Genesis Offering. Numbers are illustrative Phase 0 placeholders that will change as we ship and measure. The dashboard at localhost:4000/material-science is the source of truth for what is real today.